Failure Severity (FS) will have a direct impact on (a) Post Failure Satisfaction (PFS), (b) Satisfaction with Recovery (SWR), (c) Post Recovery Satisfaction (PRS) and (d) Loyalty.
Hypothesis H1a:
Failure Severity (FS) will have a direct impact on (a) Post Failure Satisfaction (PFS). The hypothesis was tested using ordinary least squares (OLS) regression. To examine the overall impact of FS on PFS, (PFS) was regressed against (FS). The results are presented in Table 4.30.
Table 4.30 (OLS) Regression of Post Failure Satisfaction on Failure Severity
Independent Variables Beta t
R= 0.19; R² = 0.04; Adjusted R² = 0.04; F = 15.03***
Failure Severity -0.19 -3.88***
Notes: *: significant at the p < 0.001 level, Durbin Watson: 2.09, VIF: 1.00, Tolerance: 1.00
176
Durbin-Watson statistics indicate that the assumption of independent errors is tenable. The variance inflation factor (VIF) value and tolerance statistic indicate the absence of collinearity in the data (Bowerman and O’Connell, 1990; Myers, 1990). Moreover the confidence intervals indicate that the estimates are likely to be representative of 95% of other samples (Field, 2000).
Failure Severity (FS) has a significant negative impact on Post Failure Satisfaction (PFS) as would be expected. This supports previous research in other service sectors.
Hypothesis H1b:
Failure Severity (FS) will have a direct impact on (b) Satisfaction with Recovery (SWR) The hypothesis was tested using OLS regression. To examine the overall impact of FS on SWR, (SWR) was regressed against (FS). It should also be noted that a comparison of the mean SWR and PFS figures indicate that failure severity has a greater impact on PFS than on SWR.
The results are presented in Table 4.31.
Table 4.31 (OLS) Regression of Satisfaction with Recovery (SWR) on Failure Severity (FS)
Independent Variables Beta t
R= 0.071; R² = 0.05; AdjustedR² = 0.02; F = 1.71***
Failure Severity -0.71 -1.30***
Notes: *: significant at the p < 0.001 level, Durbin Watson: 1.87, VIF: 1.00, Tolerance: 1.00
The model achieved a satisfactory level of goodness of fit in predicting the outcome variable. Durbin-Watson statistics indicate that the assumption of independent errors is tenable. The variance inflation factor (VIF) value and tolerance statistic indicate the absence of collinearity in the data.
177
Hypothesis H1c:
Failure Severity (FS) will have a direct impact on (b) Post Recovery Satisfaction (PRS) The hypothesis was tested using OLS regression. To examine the overall impact of FS on PRS, (PRS) was regressed against (FS).
The Results from the Regression test are presented in Table 4.32.
Table 4.32 (OLS) Regression of Post Recovery Satisfaction on Failure Severity
Independent Variables Beta t
R= 0.14; R² = 0.04; Adjusted R² = 0.04; F = 7.635***
Failure Severity -0.145 -2.76***
Notes: *: significant at the p < 0.001 level, Durbin Watson: 1.99, VIF: 1.00, Tolerance: 1.00
The model achieved a satisfactory level of goodness of fit in predicting the outcome variable. Durbin-Watson statistics indicate that the assumption of independent errors is tenable. The variance inflation factor (VIF) value and tolerance statistic indicate the absence of collinearity in the data (Bowerman and O’Connell, 1990; Myers, 1990). Moreover the confidence intervals indicate that the estimates are likely to be representative of 95% of other samples (Field, 2000).
178
Hypothesis H1d1:
Failure Severity (FS) will have a direct impact on (d1) Loyalty (Word of Mouth)
The hypothesis was tested using OLS regression. The results from the Regression test are presented in Table 4.33
The Results from the Regression test are presented in Table 4.33.
Table 4.33 (OLS) Regression of (d1) Loyalty (Word of Mouth) on Failure Severity
Independent Variables Beta t
R= 0.14; R² = 0.02; Adjusted R² = 0.02; F = 6.919***
Failure Severity -0.14 -2.63***
Notes: *: significant at the p < 0.001 level, Durbin Watson: 2.04, VIF: 1.00, Tolerance: 1.00
The model achieved a satisfactory level of goodness of fit in predicting the outcome variable. Durbin-Watson statistics indicate that the assumption of independent errors is tenable. The variance inflation factor (VIF) value and tolerance statistic indicate the absence of collinearity in the data (Bowerman and O’Connell, 1990; Myers, 1990). Moreover the confidence intervals indicate that the estimates are likely to be representative of 95% of other samples (Field, 2000).
179
Hypothesis H1d2:
Failure Severity (FS) will have a direct impact on (d2) Loyalty (Fly same Airline)
The hypothesis was tested using OLS regression. The results from the Regression test are presented in Table 4.34.
Table 4.34 (OLS) Regression of (d2) Loyalty (Fly same Airline – repurchase) on Failure Severity
Independent Variables Beta t
R= 0.16; R² = 0.26; Adjusted R² = 0.23; F = 9.17***
Failure Severity -0.16 -3.03***
Notes: *: significant at the p < 0.001 level, Durbin Watson: 2.00, VIF: 1.00, Tolerance: 1.00
The model achieved a satisfactory level of goodness of fit in predicting the outcome variable. Durbin-Watson statistics indicate that the assumption of independent errors is tenable. The variance inflation factor (VIF) value and tolerance statistic indicate the absence of collinearity in the data (Bowerman and O’Connell, 1990; Myers, 1990). Moreover the confidence intervals indicate that the estimates are likely to be representative of 95% of other samples (Field, 2000).
The regression however here suggests slightly higher impact of SF on repurchase than word of mouth.
180
Hypothesis H1d3:
Failure Severity (FS) will have a direct impact on (d3) Loyalty (Not switch Airline) The hypothesis was tested using OLS regression. The results from the Regression test are presented in Table 4.35
Table 4.35 (OLS) Regression of (d3) Loyalty (Not switch Airline) on Failure Severity
Independent Variables Beta t
R= 0.70; R² = 0.005; Adjusted R² = 0.002; F = 1.69***
Failure Severity -0.07 -1.30***
Notes: *: significant at the p < 0.001 level, Durbin Watson: 2.04, VIF: 1.00, Tolerance: 1.00
The model achieved a satisfactory level of goodness of fit in predicting the outcome variable. Durbin-Watson statistics indicate that the assumption of independent errors is tenable. The variance inflation factor (VIF) value and tolerance statistic indicate the absence of collinearity in the data (Bowerman and O’Connell, 1990; Myers, 1990). Moreover the confidence intervals indicate that the estimates are likely to be representative of 95% of other samples (Field, 2000).
The result show that the Beta value (-0.07) is less on switching in comparison with the Beta value (-0.16) on loyalty (repurchase) and does not have a significantly higher impact.
181
Hypothesis H1d4:
Failure Severity (FS) will have a direct impact on (d4) Loyalty (Consider this Airline my Primary choice)
The hypothesis was tested using OLS regression. The results from the Regression test are presented in Table 4.36.
Table 4.36 (OLS) Regression of (d4) Loyalty (Consider this Airline my primary choice) on Failure Severity
Independent Variables Beta t
R= 0.11; R² = 0.12; Adjusted R² = 0.10; F = 4.33***
Failure Severity -0.11 -2.08***
Notes: *: significant at the p < 0.001 level, Durbin Watson: 2.15, VIF: 1.00, Tolerance: 1.00
The model achieved a satisfactory level of goodness of fit in predicting the outcome variable. Durbin-Watson statistics indicate that the assumption of independent errors is tenable. The variance inflation factor (VIF) value and tolerance statistic indicate the absence of collinearity in the data (Bowerman and O’Connell, 1990; Myers, 1990). Moreover the confidence intervals indicate that the estimates are likely to be representative of 95% of other samples (Field, 2000).
182